1
MEDIATION WITHIN OR BETWEEN ORGANISATIONS IN FRANCE : State of development, barriers and paths
Arnaud Stimec Doctor in Management – University Paris I – Sorbonne
Paper presented at the 2001 International Association for Conflict Management conference à Cergy – Essec business school Special thanks to Andres Malave and Alan Ball for their linguistical comments on the drafts. The content remain the responsibility of the author.
2
Abstract :
Mediation, as well as the word itself, is increasingly used in France. Defining mediation as an assisted negotiation process makes it much harder to identify it as such within (or between) organisations in France. One can wonder if and why the world of organisations would be reluctant to mediation. Although many professionals express difficulties introducing mediation (and not only in France), this perspective doesn’t appear to be the main concern in the literature. Many hypotheses can nevertheless be generated from the study of mediation research and related fields: conflict, negotiation, …
The comparison of the development of mediation in France and in other countries (and particularly the US) reveals that global factors limited the development of mediation in collective bargaining. The possibility of creating an effective experience that could spread to other fields (management, business…) thus vanished. Other paths of development certainly exist. Our research on executives revealed a significant positive attitude toward mediation. Internal mediation (mostly informal management mediation) appears to be easier to accept than external mediation (mediators from mediation centres or consultants) and seems to be a reality. Factors like the perceived impact of mediation on “usual ways of operating”, relationship prospects or importance of the issue appeared to be the most significant in explaining or even predicting respondents’ choice attitude. Changing habits through training and information may certainly open new directions for mediation in organisations in France.
3
INTRODUCTION
Despite an increasing use of mediation at all levels of the French society (justice, family, neighbourhood...), mediation seems to be still largely unknown or even worse “misunderstood”. The increasing use of mediation is accompanied by an even larger use of the word, leading to a lot of confusion. For instance, travelling to Paris by train, one may well come across young SNCF mediators 1 , wearing a red and black jacket with mediation written in white, whose role is basically to help the travellers to find the right metro direction or organise the taxi queue.
In terms of organisations, there has been a significant increase in the use of processes or procedures which use the term “mediation”. Many large organisations like train or metro companies, banks, insurance companies... have their mediators. However, after further scrutiny into these services, introduced in the 90’s, we realise that, very often, it tends to refer to what is generally and internationally called an Ombudsman (corresponding here to an internal non-binding arbitration service). Defining mediation as an assisted negotiation process (for instance Moore, 1996) makes it much harder to identify it as such within (or between) organisations in France. In fact, its formal and effective existence seems to be an exception for the observer; despite the creation of several mediation centres in the last decade or some original experiments within a few organisations (Stimec, 1999).
The gap in France between mediation in the business world and mediation in society, or with organisational mediation2 development in some other countries, leads us to ask a few questions and to work along two research paths. The first path is to explain the past and present situation from a political, historical and institutional perspective (Part II) , in particular in industrial relations settings. What is the real state of the development of organisational mediation in France ? What forms does mediation take ? What explains the gap with other types of mediation in France or with other countries ? Is there specific resistance to mediation in France in organisational settings ?
1
They were hired after a large government subsidised campaign to help young men and women under 25 find a job. 2 By this expression, we wish to encompass mediation inside organisations (collective bargaining, daily working relationships,…) or outside (commercial partners of all sorts, other organisations).
4
The second path is to explore the future prospects for mediation in French firms3 (Part III). Is there a need for mediation ? What form of mediation can we expect in French firms ? What factors make mediation more expectable ? We will start with a review of the literature on resistance and barriers to mediation (Part I).
3
The firm being a specific type of organisation that doesn’t allow generalisation to all organisations. Based on the specialised literature (see Rojot & Bergman, 1991), we define organisations as social constructs with goals and regulations. The firm is an organisation whose main goal is profit.
5
PART I - RESISTANCE AND BARRIERS TO MEDIATION THROUGHOUT THE LITTERATURE
One initial perspective to answer the previous questions could be to explore resistance and barriers to mediation identified in various settings and from various levels of analysis. Although some research or publications refer to this problematic, this does not seem to be a major issue in the literature. Nevertheless an essential point for professionals is : how to interest people in taking part in mediation (especially the other party)? How to develop the recourse to mediation services? We thus decided to explore the mediation through the literature and related subjects (conflict, negotiation, power...) with an open view in order to produce as many hypotheses as possible. We will close this section with a mention of existing research and other writings.
The first entry lies in the meaning of mediation itself. When we examine the possible meanings of mediation in France : information, broadcasting, non-binding arbitration, some forms of pedagogy... and compare them to an albeit very broad definition such as “assisted negotiation”, we realise that this must be very confusing for the non specialist. This confusion may be aggravated by some psycho-sociological phenomena such as the inability to consider the presence of a third party (Freund, 1983), the fact that conflict tends to make involved parties think in a binary way. One may be with or against but not neutral. This confusing diversity and the effect of conflict on the perception of third parties may thus constitute some form of barrier or resistance toward mediation.
Conflict being one of the favoured grounds for mediation, delving into conflict theory should also reveal some prevalent attitudes. Conflict is a subject of controversy in itself or is at least considered with different lenses according to the author’s field of study. A psychologist may underline conflict as the result of a different perception, frustration or an impulse (For instance Thomas, 1979 ; Touzard, 1977 ; Rubin & al., 1994). Sociologists or anthropologists may insist on global factors leading to conflict or aimed to handle it (For instance Coser, 1982 ; Girard, 1972 ; Simmel 1995). Jurists may treat conflict as a question of application or applicability of law. Organisational theory specialists may perceive conflict as an obstacle to
6
decision making or efficiency4 ... We could multiply the points of view on conflict. An extended review of them produced the following observations.
First of all, conflict is, most of the time, considered in organisational theory as a negative thing (Thomas, 1992) despite some original contributions that take the opposite view (De Dreu & Van de Vliert, 1997 or Pascale, 1991). Secondly, mediation is no more than one process among others and is not always a possible choice. Thirdly, conflict escalation (Rubin & al., 1994) tends to reinforce the limited rationality of the actor (March & Simon, 1974) and thus may limit the recourse to mediation (as it is asserted in Zubek & al., 1992). Broadly speaking, the perception of an unacceptable deadlock or of superordinate goals may create a favourable context for mediation. On the contrary, the risk of loss of face or power tend to be barriers to mediation (Arrow & al., 1995).
Power is another very important dimension of the environment of mediations. This paper is not the place to describe controversies about its real nature and definition (see for instance Rojot & Bergman, 1991). Defining power broadly as “the possibility for some individuals or groups to act upon other individuals or groups” (Crozier & Friedberg, 1977, p.55 et ss.) opens the path for two basic effects in negotiation : push to do something or prevent from doing something (Dupont, 1994). Some of the identified properties known about power may help to extend our view on barriers to mediation.
It is first asserted that power is closely associated with negotiation (Rojot, 1994). There would be no negotiation (nor mediation) if there was no power to give, take, inform, explain or consider someone... Power and negotiation rely on two basic mechanisms of influence or exchange : uncertainty (Crozier & Friedberg, 1977) and mutual interdependency (Bacharach & Lawler, 1982). The power of one party is also subjective (interpretation, selective perception of the environment), contingent (environment, parties, issues) and independent of the other. This leads us to consider the perceived power balance in a particular setting as an influence in the recourse to mediation. It is generally asserted that power imbalance gives a better negotiation outcome to the stronger, although this may be questioned (Zartman, 1997). This is often difficult to measure in a non-tautological way (i.e. analysing who has got the bigger slice). Power balance may for instance be measured by what is very often called the Batna (best alternative to a negotiated agreement) [Fisher & Ury, 1982]. A strong Batna of 4
See Rojot (1994 & 1991) for references and detailed presentation
7
one the parties involved could make mediation less amenable than a weak Batna. Moreover, the struggle to get the stronger power position may be, as it is for negotiation, a barrier to mediation, by shaping rigid attitudes and a non genuine negotiation setting.
Considering negotiation in general, some other factors are supposed to play a role in the choice to negotiate : the interest in cultivating a long term relationship (Fisher & Ury, 1982 ; Zubek, 1992) or the positive or negative influence of other people (for instance colleagues, subordinate or supervisors) [Rojot, 1994 ; Dupont, 1994]. We presume they also may affect the choice to mediate or not.
Research and publications referring explicitly to barriers or resistance to mediation are not so common. We will start by considering broad barriers to conflict resolution or mediation, non specific to organisational settings and progressively come around to our theme.
Considering barriers, Mnookin (in Arrow et al., 1995, p.2) suggests there are three types : strategic (for instance studied in game theory), psychological (for instance limited rationality) or structural. A mediation manual (Moore, 1996, p.83) highlights four causes for mediation refusal : unfamiliarity with the process, preference for win/lose situations, strong emotions or the habit of using judicial means of settlement. A French Professor (Salzer, 1995) identified 21 types of resistance and refusal to mediation and classified them into three categories similar to the one above : unfamiliarity with the process, doubts over the process, preference for another conflict management system. Based on this work and some other research in Victim-Offender mediation in Germany (Netzig, 1996), we proposed a re-framed model of motives and resistance to Victim-Offender mediation in France (Stimec, 1997), which derive from attitudes toward facts, relationships, psychological perceptions, expected results, the mediation process and the meaning of mediation. Studies on lawyers’ attitudes in the UK (Genn, 1998) revealed three types of resistance (ignorance or lack of experience in mediation, a fear of perceived weakness by the other side, a fear of compromise) and two main motives (saving time and money ).
International settings (Bercovitch, 1996 ; Raiffa, 1995, p.144 ; Susskind & Babbit, 1992, p.30), may help us to expand the hypotheses, although some barriers or incentives to mediation may be very specific to the context. This points out the importance of the mutual recognition of parties and their authority as speakers, the contingency of themes (territorial
8
matters seems to constitute stronger barriers) and the personality or reputation of the mediator. Some other factors were indicated earlier in this article such as the role of deadlocks, the pressure to mediate, or some practical considerations (neutrality, cost, ...).
The development of Environmental mediation interests us because it very often includes multiple levels of actors and organisations. Although the future of environmental mediation appears to be limited in France, due in particular to the numerous political negotiation levels that often pretend to be a mediation space (Mermet in Faure & al., 1998), the difficulties faced in other countries may provide hypotheses for our research. One paper has caught our attention and concerns the attitudes of environmental actors toward the possible recourse to mediation (O’laughlin & Schumaker, 1998). The authors identify five levels of resistance : fear of interference with the managers’ authority, doubts about the efficiency of ADR, fear of interference with scientific decision making, difficulty to include all environmental groups, lack of perceived legal authority to use ADR.
Organisational mediation seems to be the subject with the least publications than any other setting (family, school, court…) [Folger & Jones, 1994 ; Kressel & al., 1986]. The study of resistance is rarely treated and is often secondary. Through the related experience of professionals or action-research (Costantino & Merchant, 1995; Deleon, 1994, p.70 ; Susskind, 1991 ; Ury, Brett & Goldberg, 1988) we have confirmation of a general resistance observed in other fields such as : a concern for neutrality, expertise of the third-party, intentional conflict, power asymmetry, preferable non-negotiation alternatives, confidentiality violation (from the mediator or parties), need for jurisprudence or a recognisable winner, fear of concessions or ignorance of ADR. Other factors, such as culture (or uses), perceived weakening of authority or symbolic perception of failure as a manager may appear to be more specific to an organisational context.
An extended survey of the top 1000 U.S. firms (Lipsky & Seeber, 1998) revealed a strong motivation for mediation in relation to saving money and time, third party expertise, preservation of the relation, or simply when mediation is required by contract or judge. Some significant types of resistance include fear of compromise, doubts about the mediator (expertise, confidentiality, neutrality…), hierarchy refusal, the legal value of the agreement or the strategic risk.
9
Another opportunity to find hypotheses in the literature was to examine research on mediation success (see Henderson, 1996 for a complete review). Although this is a different problematic than ours, we may assume that contingent factors in favour of mediation success or leading to more mediation failures may also constitute incentives or barriers to the recourse of mediation. Unfortunately, it seems that there is no agreement on the relevant factors and whether they play a role. Many other factors may influence the outcome. For instance, a “dealmaker” or “orchestrator” style 5 may be more or less efficient according to conflict intensity. It is thus risky to assert that conflict intensity may be a barrier to mediation in general. Brett, Barsness & Goldberg (1996) nonetheless assert that two contingent factors should be taken into account : when one party has no financial interest to mediate or when there is a possibility (even low) of high gain elsewhere (Jackpot syndrome).
Most of the publications do not provide a statistical correlation with recourse to mediation decision making or attitudes to mediation in face of a dispute or conflict. It is thus very difficult to know which of these factors appear to be decisive; they may certainly contribute to explaining some particular refusals or non-occurrence of mediation but are not specific enough to respond to the current state of affairs: a weak recourse to mediation in or among organisations in France. To go deeper into the field, we will search for global institutional factors (part II) and explore key factors of the attitudes towards mediation at the level of the actor (part III).
5
See Kolb (1983).
10
PART II - ORGANIZATIONAL MEDIATION IN FRANCE AND ELSEWHERE
Despite the same theoretical promises, mediation hasn’t developed in the same way throughout the world. Professionals very often assert that mediation is a hard case in organisations in France because of our culture. Comparing the state of development of mediation in France, Italy… and North America or even Great-Britain suggests that Latin countries (or catholic) may be structurally less open to mediation than Anglo-Saxon countries (or protestant). We think that this type of cultural explanation, whether valid or not, doesn’t give much insight into the reasons for this difference (Zartman, 1993). What, in our culture, may be a real barrier to mediation ? Could it be a pyramidal view of social relations (Demorgon, 1996) ? But then, why do we have this trait, what maintains it ? Considering culture as cause and consequence of social situations (Faure & Rubin, 1993, p.227) leads us to consider the perception of mediation as having a structural6 cultural nature (largely the fruit of history) and being the result of limited but rational choices anchored in the present (March & Simon, 1974).
The review of the development of mediation in the United States, leads us to the following conclusions and to a model constitutive of an explicative framework.
1) Despite some earlier experiments, mediation did not play a significant role until the second world war (War Labour Board) to help channel social unrest in industry, at a time when the cost of strikes and their national strategic consequences were perceived as unacceptable (Dunlop & Zack, 1997; Necheles-Jansyn, 1990). Agreements leading to some form of union recognition (and sometimes mandatory representation) in exchange for peace agreements (where strikes were forbidden). Mediators were sent and even imposed by the government in order to prevent strikes. This led, in 1947, to the constitution of the FMCS (Federal Mediation and Conciliation Services) and later on to state mediation agencies. Thousands of cases were treated by both agencies.
2) At the same time, legal recognition of collective agreements and arbitration processes led to more and more agreements including arbitration. Mediation often emerged as another chance before arbitration. This spread at a collective level to the interpretation of agreements (rights disputes) but also during re-negotiations (interest conflicts). Arbitration
11
and mediation later appeared as part of various grievance processes (Sloane & Whitney, 1972; Ury, Brett & Goldberg, 1988). Arbitration seems to be more acceptable than a strike or less costly than endless judiciary procedures. Mediation appears to be the last chance for mutual acceptable agreement before arbitration.
3) Mediation progressively spread into different other areas as a strategic function (Costantino & Merchant, 1995) for working relations as Ombuds (Rowe, 1991) or even in management with the manager as mediator (Carnevale, 1986; Cohen, 1999; Karambaya & Brett, 1994; Kochan & Bazerman, 1986; Kolb, 1986; Nugent, 1997). Mediation followed a limited but similar development in the public sector (Dilts & Walsh, 1988; Metzger, 1990; Prasow & Peters, 1970), where strikes are often forbidden in exchange for arbitration and mediation procedures. We argue that the development of mediation in other areas through this experience followed a similar pattern. For instance, in the commercial area, disputes among firms found in arbitration a faster and cheaper solution and later on, mediation appeared as potentially even more satisfactory (arbitration becoming sometimes a complicated procedure). Although we have no strong evidence, we could even argue that this fostered the development of mediation in private life (family mediation, neighbourhood mediation…).
Considering this model, mediation in France, although probably the second oldest profession7 and despite many attempts, has been historically limited in collective relations 8 by several factors. One significant period was between 1936 (where major strikes spread throughout the country) and 1938 (Didry, 1998; Naulin, 1998) : 12 477 cases out of 12 939 were solved through arbitration and mediation, following the legislation of this time. We argue that the compulsory and effective form of arbitration was, at the time, a real incentive to mediation9 as it has been in the United States. Due to the recognition of the right to strike and the historical role of various unions (participation in the liberation of France), the subsequent legal attempts to introduce mediation or arbitration always failed (Javillier, 1981; Naulin, 1998).
6
See A. Giddens (1984) a theory of structure. It’s feminin « médiatrice » appeared in 1611 according to the « Petit Robert ». 8 See (Touzard, 1977 ; Adam & Reynaud, 1978 ; Amadieu, 1999) for details on French industrial relations. 9 Mediation was at that time often called conciliation, as it was in the USA. We don’t want to discuss the differences between mediation and conciliation. We would like to consider both of them as a whole, recognizing the existence of different mediation styles and goals. 7
12
For instance, the 1950 procedure included a mandatory conciliation but optional arbitration (thus with no pressure to agree). The 1982 procedure (Loi Auroux) appears largely unused and even the few conflicts where a national mediator is chosen do not refer to it (Salzer, 1999). Other forms of informal mediation, for instance the laboratory inspector (Touzard, 1999) play a role in some occasions.
As for individual labor disputes, there is an exclusive authority of the “Conseil des Prud’hommes” and equal representation labour court where judges are elected by workers and employers. This court has been in existence since 1806 and initially almost exclusively functioned by conciliation (Cam, 1981). Conciliation is mandatory and followed in case of failure, by a decision from the elected judges. The conciliation rate unfortunately declined (below 10% at national level). This example could appear to be a refute of our explicative framework. Our further investigations (a case study in Paris and Nantes 10 ) among different units lead to the following observations : -
Often, only 10 minutes are given to conciliate (better results are observed in the units where more time is given);
-
No training or poor training is given to negotiation, mediation or conciliation techniques (a video training tool we obtained, which is supposed to show how it should be done, shows the failure of the conciliation after the initial proposal of the elected judges- which is essentially a 50/50 compromise attempt) ;
-
The same persons will judge the case if conciliation is to fail (knowing this, parties may limit what they say) ;
-
One of the parties (the employer most of the time) is very often not present and represented by a lawyer whose mandate is not to conciliate ;
-
Some parties may have an interest to make it last, waiting for appeal, as no penalty may be applied, thus gaining several years.
A recent experience in the Grenoble appeal court shows a success rate of 60% (when mediation is accepted) although the cases had failed in conciliation and were judged at the “Conseil des Prud’hommes” (Blohorn-Brenneur, 2001). This clearly confirms that opportunities are missed at the level of the “Conseil des Prud’hommes”.
10
We triangulated different data proceeding with interviews of conciliators, interviews of parties, observation of conciliations, file analysis… following the methodology of Wacheux (1996) and Yin (1984).
13
We think that mediation is less developed in France than in some other countries like the US, because the opportunity to develop it into an effective social tool were missed. In particular, we tried to show that collective (see below) or individual labour relations couldn’t provide the social apprenticeship needed. Examination of Canada or Great-Britain confirmed the pattern observed for the United States.
France11
U.S. Mediation
emerged
from
collective Mediation
bargaining because : Federal
government
encouraged
mediation
could
not
spread
in
the
collective bargaining because of : and
State
(pressure
strongly The French government played a strong role on
the in the issues (pressure on the content), often
process)
deciding for the parties 12
There is often one representative union
There are (by law) five representative unions that may never all agree
Unions can commit to peace agreements The right to strike is part of the French including strike limitations
Constitution and no-one can commit to a no strike agreement (this would be illegal)
Arbitration is legally recognised (Supreme Arbitration is forbidden by law and is Court, Steelworker Trilogy, 1960) and creates restricted
to
certain
areas
(commercial
a strong source of development for mediation relations for instance). The fact that the development of mediation followed a certain path in some countries doesn’t prove this is the only possible one. In fact, many efforts have been made in France by various organisations or individuals to promote mediation. Despite difficulties to make it significant within or between organisations, some developments may be expected. Based on the literature on the motives and resistance toward mediation, and on the result of the comparison presented above, we then tried to understand why mediation is still largely unused within or between organisations and what are the possible prospects for it.
11
We don’t assert that one system is better or worse. Our purpose is simply to explain what conditions created a favorable context for mediation. 12 Political science and particularly the comparison of French and American revolution help to understand the choice towards more liberalism (and federalism) on the one hand and welfare state (and centralism) on the other hand. One can also compare the size of the codes about working relations in France and USA.
14
PART III – EXPLORING THE PROSPECTS FOR MEDIATION IN FRENCH FIRMS
Based on the numerous possible theories which explain the recourse to mediation, we proceeded with an exploratory survey to identify the variables useful for our purpose. Questionnaires with open questions, interviews with professionals and file analysis helped us to triangulate the data (Wacheux, 1996). This exploratory survey, the literature (part I), the comparison with various fields and countries (part II) and our experience as mediator, led us to formulate hypotheses.
Hypotheses
Although mediation is largely unknown and unused in organisations, we assume that in general attitudes are positive (H1.1). If faced with a precise setting, respondents may nevertheless be less in favour of mediation (H1.2). The gap between this general attitude and attitudes in-context or the real behaviour (weak use of mediation) is one basic legitimacy of this research. One original result from the exploratory survey was the different perception of internal mediators (anyone within the organisation recognised as playing this role whether formally or not) and external mediators13 (consultants, mediators from mediation centres…). We think parties may be more open to internal mediators than external ones (H2.1). All other things being equal, external matters (business relations for instance) may be more favourable to external mediation and internal matters to internal mediation (H2.2).
Some general factors may play an active role in the different attitudes (general attitude, attitudes toward external mediation and internal mediation based on respondents cases): the respondent’s relation to conflict or negotiation (H3.1), usual way of operating in the organisation (H3.2), social groups defined by sex, age, qualifications… (H3.3) or the preference for management arbitration as a means of ruling internal conflicts (H3.4).
Other factors are contingent to the particular conflict situation: conflict escalation (H4.1), deadlock (H4.2), alternative (H4.3), attitude toward negotiation in the particular situation (preference, perception of feasibility) [H4.5], type of conflicts (H2.2), confidentiality
13
We will use the expressions internal mediation and external mediation to design these levels of intervention.
15
(existence of confidential issues, perception of the mediators respect for it, concern for breach of confidence) [H.4.6].
The level of knowledge or apprenticeship of mediation is a key factor and particularly the perception of self when using mediation (H5.1), the correct understanding of mediation (H5.2), familiarity with the concept (H5.3). We argue that confrontation with mediation in particular cases creates a positive and affirmative attitude (H5.4).
Issues or perceived interests in the particular situation may also influence attitudes. It may be the perceived consequences on career (H6.1), the perceived time needed to treat the case (H6.2), the financial issue (H6.3), the long-term relation with the other party (existent or desired) [H6.4] or the desire to use mediation for a strategic purpose (H6.5).
Finally, a decisive factor may be top management’s impulse in favour of mediation (H7.1). We think there is a dramatic lack in this support. Unfortunately, top management may be resistant to mediation for fear of losing authority (H7.2) but may prefer internal mediation as respectful of internal order (H7.3).
Population and Methodology
Based on the various hypotheses, a questionnaire has been developed and re-read by several professionals or researchers. The questionnaire is comprised of two parts. The first part measures various general attitudes on mediation, conflict or other related subjects. The second part requires the participant to refer to their most striking personal conflict (or difficult negotiation) at the office faced in the last year. In the second part, many items help to identify attitudes and link them to an agreement (if it was proposed) to mediate with an internal mediator (formal or not) or with an external mediator. Because of limited space, the questionnaire (8 pages) is not presented here. In the results section, significant items are listed and detailed.
Because of the number of hypotheses and the small amount of research on the subject, we had to represent each variable with only a few items (and often only one), sacrificing an optimal measure quality. The questionnaire is thus exploratory. It was administered between Nov.1999 and Dec. 2000 to 315 managers participating in a two or three day negotiation
16
workshop (considered as vocational training). The response rate was 83% thanks to the collection procedure (the questionnaires were given one day and collected the following, or answered immediately in about 20 minutes).
Women constitute 35% of the sample. This is a little bit higher than national French data of women working at this management level (29 %). Most of them were taking part in such a workshop for the first time (85 %) but a few were returning to continue their personal training (other negotiation workshops : 13 % or mediation workshops 2 %). A large part of them considered themselves as executives (71 %) and some as top executives (19%).
We used a Mann-Whitney U Test to test the dependency of various variables (ordinal Likert items scale) with binary (yes/no) attitudes towards the recourse of a mediator (internal or external). We only considered variables presenting a dependency significant at a 0,05 level or better. When considering the dependency between two ordinal items, we simply used correlation coefficients.
Logistic Regression is a second step in data analysis. Classical regression is not suitable for binary or dummy variables. Introducing a probabilistic formulation and a Logit make regression possible and meaningful (Menard, 1995; Pampel, 2000). This will enable us to propose a predictive model of attitudes toward mediation and at the same time underline the most significant variables.
Results
Frome the data, we can get a picture of conflicts in French firms. Most respondents think that relations with colleagues are not conflictual in general (over 80 %) but regard relations with commercial partners (28 %) or unions (58 %) as generally conflictual. Examining the respondents cases (part two) show that most conflict situations are part of everyday management (45 %, i.e. project management, decision making…) or commercial relations (30 %, i.e. clients or suppliers). Most situations present a weak (less than 7000 Euros) or null financial issue (56 %) but high level issues (more than 76,000 Euros) are still important (23,4 %). Respondents generally think their case is worth spending at least half day trying to solve it (83 %). About the same level of participants felt they were in a deadlock and that negotiation was their favoured settlement process.
17
The general attitude to mediation is positive (agreement or strong agreement that mediation is a good thing for organisational conflicts for 79 % of the respondents (H1.1). When asked the same question in reference to a particular context (for example a conflict with a colleague, a conflict with another firm…) results are all lower than this (H1.2). Participant’s cases confirm tendencies with a result in favour of mediation which is lower again.
(results expressed in %) Attitudes towards mediation in a situation including : 1 – two subordinates External Mediator
(in general) YES NO 18 82
(particip. cases) YES NO 28,6 71,4
Internal Mediator
74,4
22,6
62,5
37,5
External Mediator
35,5
64,5
22,2
77,8
Internal Mediator
76
24
60
40
External Mediator
47,2
53,8
42,4
57,6
Internal Mediator
71
29
64,7
35,3
4 – respondent and colleague of same level in the hierarchy
External Mediator
36,4
63,6
32
68
Internal Mediator
88
12
70
30
5 – respondent and commercial partner
External Mediator
70,7
29,3
37,2
62,8
Internal Mediator
68
32
52
48
2 – respondent and a subordinate 3 – respondent and supervisor
Internal mediation is preferred in all types of situations. The external mediator seems to be unwelcome when intervening in relation to the respondents’ power position (conflicts including two of his subordinates or the respondent and a subordinate). External mediation seems to be more acceptable when facing an external relation or with supervisors. We think that an internal mediator would be hard to accept for the other party when it involves a commercial partner (about 70 % of respondents think so), and this anticipation may explain the lower result of internal mediation (52 %). This data must be compared to the 5% of all cases where an external mediator (or assimilated) was required (on average 36,6 % of participants would have agreed to external mediation) and 28,8 % of situations where an internal mediator intervened (compared to an average of 60,9 % who would have agreed to it). Many opportunities for mediation may thus be lost.
18
We now will try to understand what variables have an impact on the attitudes described ahead. Due to the number of variables tested and the limited space, we will only present data with significant relations to the attitudes. The tabs below present the results. We coded (+) when stating the variable tending to be in favour of YES. We indicate the degree of statistical significance by (*) for 0,05 and (**) for 0,01. A blank means the relation was not significant or not applicable. The related hypothesis number is indicated in between [ ] after the item name and formulation.
Description of the variable
General Attitude
External Mediation
GENERAL ATTITUDES – FIRST PART Sex = woman [H3]
(*) ; +
Age = younger [H3]
(*) ; +
(*) ; +
Use = mediation would in general disturb the usual way of operating (asked in the first part) [H3] Negotiation = like to negotiate [H3]
(**) ; -
(**) ; -
Conflict = participant feels deep relief when a disagreement is settled [H3] Familiarity with mediation = heard of mediation in organisations (1) [H5] Familiarity with mediation = to have heard that mediations took place in my organisation (2)
(*) ; +
Internal Mediation (*) ; +
(**) ; +
(*) ; + (**) ; +
19
CONTEXTUAL ATTITUDES IN REFERENCE TO THE PARTICIPANT’S CASE – 2 nd PART Use = mediation would inconvenience the usual way of operating in this case [H3] Colleagues = would have approved [H4]
General Attitude
External Mediation (**) ; -
Internal Mediation (**) ; -
(*) ; +
(**) ; + (*) ; -
Escalation = relation is perceived as hostile or tense rather than serene [H4] Deadlock (1) = perception that there was a deadlock [H4] Deadlock (2) = perception that the situation is difficult [H4] Negotiation = perception as possible [H4]
(*) ; + (**) ; + (*) ; +
Confidentiality = rupture of confidentiality is perceived as a trouble in the context [H4] Neutrality = mediator is impartial [H5] Image (1) = recourse to mediation shows good will [H5] Image (2) = is a personal failure
(*) ; (*) ; + (**) ; +
(*) ; +
(*) ; -
(**) ; -
Image (3) = recourse to mediation shows good faith
(*) ; +
Image (4) = recourse to mediation shows weakness
(*) ; (**) ; -
(**) ; -
Excessive formalisation = mediation perceived as leading to… [H5] Time = the situation is worth spending at least half a day trying to solve it [H6] Motivation (1) = gaining time [H6]
(**) ; + (**) ; +
(**) ; +
Motivation (2) = preserving the relation
(**) ; +
(**) ; +
Relation = will to cease the relation even in case of agreement [H6] Strategy = mediation recourse would show good will in case of management arbitration [H6] Financial issue = when increasing [H6]
(**) ; -
Cost (1) = the cost of an external mediator would have appeared to be too important Cost (2) = it was worth paying somebody to help in solving the problem Supervisor = would have approved [H7]
(**) ; +
(*) ; +
(**) ; +
(**) ; -
(**) ; (**) ; + (**) ; +
No significant relation was proven for other variables. These results tend to refute some of the hypotheses and temporarily confirm others. Hypothesis 1, as it was shown earlier, is confirmed by the results. Hypothesis 2 is partly validated but there is a lack of statistical significance to assert contingency.
20
Regarding hypothesis 3, the “usual way of operating”, the fact that participant like to negotiate or the age of the respondent appeared to be influential on the attitude in choosing external mediation. Only sex appeared to be significant at this level for internal mediation. The perception of “usual ways of operating” might be influential for internal mediation but this is only verified for the contextual variable. This may mean that this perception is not effective as an a priori but rather in situ. Moreover, we think internal mediation appears as less unusual than external mediation (54 % of respondents think external mediation would be a challenge to usual ways of operating compared with only 30 % only for internal mediation).
Hypothesis 4 is partly verified concerning internal mediation only. The role of deadlock, perception of negotiation as being possible and conflict escalation is congruent with the literature. The absence of the influence of deadlock and perceived alternatives on the attitude in choosing an external mediator is surprising. Nevertheless, we verified this was not due principally to a quality of measurement by testing the correlation between the perception of negotiation as possible or preferable and this variable. This was consistent with the literature showing deadlock as being neutral regarding preference to negotiation but that a good alternative was not encouraging for negotiation. The measure of the conflict escalation appeared to be poor and we can draw no conclusion14 . Conflict types did not significantly influence the preference although data showed a significant difference between internal mediation and external mediation. Further investigations are needed at this level. Colleagues attitudes and the existence of confidential issues appeared to influence the attitude toward internal mediation.
For hypothesis 5, familiarity with mediation has proven significant concerning internal mediation. Variables relating to the respondents’ image and the risk of excessive formalisation were significant regarding the recourse to mediation but may mainly be a confirmation of the contextual choice. We also checked the understanding of mediation and that no blockage was due to it. No structural experiential learning (the effect of having experienced real mediation on general attitudes regarding mediation) was diagnosed although more data would be needed at this level.
14
We think the measure of escalation a couple of months after a case with a questionnaire hardly an appropriate tool.
21
Hypothesis 6 is largely confirmed. The time available for the issue, the financial stake, and the long term relationship attitude are highly (external mediation) or significantly (internal mediation) associated with the choice attitudes. Motivation for time sparing or showing good faith to the hierarchy proved significant. Career prospects didn’t seem to be relevant here.
No significant influence of hierarchical position was confirmed (hypothesis 7). On the contrary, higher rank managers asserted a more open attitude to mediation. Resistance may be higher at the middle level management that could perceive mediation as a loss of face. Contrary to our hypothesis, top managers seem more open than other managers toward external mediation (although internal mediation remains preferred).
Logistic regression modelling
A second step in the analysis is to check which variables are the most effective to predict or explain the decisions. Logistic regression, based on the variables identified as previously significant helped us to build an optimised model (predict or explain most decisions with as few variables as possible). Before proceeding to logistic regression, we also eliminated items that may be a rationalisation or may not be predictive (for instance : an internal mediator can better understand the situation). Moreover, concerning attitudes measuring the perception of mediation on the usual ways of operating, we preferred those measured in the first part (the respondents could not be influenced by their own choice in responding, because this is before a personal case was chosen). We used SPSS 7.5 statistics software for the various proceedings.
Recourse to external mediation
Logistic regression treatments lead to a couple of variables being selected for each of the two choice attitudes. The desire to recourse to external mediation is mainly explained by four predictive variable. The first one is a general attitude in the first part of the questionnaire : the “usual way of operating ” (whether mediation is perceived as inconvenient to it). The three others are contextual : “strategic attitudes” (mediation is valued as showing good will), cost (it was worth paying somebody to help solve the problem) and relation (future relation is expected). The model correctly predicts nearly 76 % of choice attitude in relation to external mediation, is recognised as statistically significant (sig. = 0.000) and has a pseudo R² of 0,2
22
(the Nagelkerke R², aimed to favour comparison with classic regression indicates 0,317). Applying the model with its coefficients to two small samples collected differently (20 members of a negotiation club and 19 persons working in firms at managerial level, whose email address was given by people of our personal network) shows respectively 74 % and 80 % good predictions. This shows the good predictive and explicative capacity of the models (both better than 50 % and better than the percentage of yes and no). The coefficient B shows the respective importance of each variable (to compare in absolute value). The financial issue (coût3) is the more influential. Other variables are at a comparable level.
---------------------- Variables in the Equation ----------------------Variable
B
S.E.
Wald
df
Sig
R
Exp(B)
UTSTRAT1 INTCOM2 NCUSA1E COÛT3 Constant
,3616 ,2768 -,3524 ,8194 -2,6092
,1542 ,1317 ,1412 ,1469 ,9640
5,4973 4,4168 6,2279 31,1352 7,3253
1 1 1 1 1
,0190 ,0356 ,0126 ,0000 ,0068
,1001 ,0832 -,1100 ,2888
1,4356 1,3188 ,7030 2,2692
A logistic regression equation of the model is :
L = 0.3616*utstrat1 +0.2768*intcom2 -0.3524*ncusa1e +0.8194*coût3 -2.6092 and P = exp (L) / (1 + exp (L)) If P>0,5 then the prediction is « no ». If P
